EFFECTS OF REED CANARY GRASS (Phalaris arundinacea) ON TERRESTRIAL ARTHROPOD 
ABUNDANCE, BIOMASS, AND DIVERSITY IN UPPER MIDWESTERN RIPARIAN WET MEADOWS.

Melissa S. Meier 1,2, Eileen M. Kirsch1, and Robin Tyser2   
1U.S. Geological Survey, Upper Midwest Environmental Sciences Center, La Crosse, 
WI 54602.  2River Studies Center, Department of Biology, University of 
Wisconsin-La Crosse, La Crosse, WI 54601. 

The invasive nature of reed canary grass (Phalaris arundinacea) is widely known.  
However, effects of reed canary grass (RCG) invasion upon wildlife are not 
known.  Current research indicates that wet meadow birds do not seem to be 
negatively affected by RCG.  However, for reed canary grass to provide suitable 
habitat for birds, it must provide adequate resources for nesting and foraging.  
Arthropods are the main source of food for many grassland birds, especially 
during the nesting season.  We hypothesized that sample biomass (wet weight, in 
grams), number of individuals, and diversity of the terrestrial arthropod 
community vary in relation to RCG dominance (measured as percent cover) at 
sample points.  Twelve wet meadow sites were selected in SE Minnesota and SW 
Wisconsin, representing a range of RCG infestation.  Study sites ranged in size 
from 4.5-16.5 ha.  Each site was marked with PVC poles at    50-m intervals to 
create a grid.  At every point of the grid, all plant species within a 2-m 
radius were identified and percent cover of each species was estimated.  Maximum 
vegetation height, vegetation height density and litter depth were also 
recorded.  Arthropod sampling was conducted in late June-early July in 2001 and 
2002. This time period approximately coincides with peak chick rearing and may 
represent arthropod species available to birds during this critical stage.  
Arthropods were sampled with sweep-nets at 3 to 8 randomly selected grid points 
at each plot (73 samples per year).  A sweep net sample consisted of walking a 
20-m transect in a random direction, making upward and downward sweeps through 
the vegetation, alternating sides with each step.  Net contents were placed in a 
killing jar containing a few drops of ether then transferred to a labeled sample 
jar.  Arthropods were separated from plant material and counted.  The entire 
arthropod sample was weighed to the nearest 0.0001-g and preserved in 70% 
alcohol.  Samples were sorted to order and identified to family.  We modeled the 
effects of vegetation variables on log (biomass) and counts of individuals per 
sample and counts of families per sample.  Variables included in the analyses 
were percent cover of RCG, percent cover of forbs, average vegetation height 
density, and average litter depth.  Eight a priori models were considered.  
Biomass models were fit using a linear mixed procedure (PROC MIXED, SAS 8.1) and 
the models for counts of individuals and families per sample were fit using 
generalized linear procedures (PROC GENMOD, SAS 8.1).  Because this analysis was 
exploratory, models were averaged to reduce model selection bias.  Percent cover 
of RCG, percent cover of forbs, and average litter depth were not associated 
with log (biomass). Vegetation height density was negatively associated with log 
(biomass).  Percent cover of forbs was positively associated with number of 
individuals and number of families per sample.  The other vegetation variables 
were not associated with either count variables.  The most abundant orders were 
Diptera, Homoptera, and Hymenoptera, followed by Hemiptera, Coleoptera, and 
Araneae.  There also appeared to be differences in family composition between 
points of low RCG and high RCG.  Although data analysis is still in progress, we 
have not detected a clear relationship between RCG and the arthropod community.

Keywords: Phalaris arundinacea, reed canary grass, terrestrial arthropods, 
vegetation height density, wet meadows